Research And Implementation Of Network Content Security Audit Platform Based On Stream Computing

With all the people entering the Internet Age, information security is becoming more and more important. When people focus on the network attack, but ignore the network security based on content. The network security based on content is the semantic level of information security, it is a direct layer impacting on the public users. So, it is particularly important to this aspect of research.Analyzing the disadvantages of traditional network content security audit scheme based on matching rule. In vectoring records, combined with the spliting words detection technology, the homophone matching technology and dynamic characteristics of the right value calculation method, solves the traditional scheme which can not analyze spliting words, homophone. Using the knowledge of machine learning, solves the reversed order problem and deep learning problem. At the same time, the incremental SVM(Support Vector Machine) learning scheme based on K-means is proposed, which can speed up the learning efficiency while ensuring the accuracy of the traditional support vector machine.Finally, the entire platform architecture in the stream computing, to face the plight of network security data environment.The improved scheme can effectively deal with the problems of high concurrency,large capacity and other issues in network information security, and can solve the disadvantages of traditional network content security audit scheme. In machine learning,both accuracy and speed can be improved. Therefore, the network content security audit still has a long research significance.